Malware is the main threat for all computing environments. It also acts as launching platform for many other cyber threats. Traditional malware detection system is not able to detect “modern”, “unknown” and “zero-day” malware. Recent developments in computing hardware and machine learning techniques have emerged as alternative solution for malware detection. The efficiency of any machine learning algorithm depends on the features extracted from the dataset. Various types of features are extracted and being researched with machine learning approach to detect malware that are targeted towards computing environments. In this work we have organized and summarized different feature types used to detect malware. This work will direct future researchers and industry to make decision on feature type selection regarding chosen computing environment for building an accurate malware classifier.